Language Log

Recent advances in machine learning have put an emphasis on the role of big data in studying language structure and linguistic diversity. However, the languages for which there are big data represent only a very small amount of the range of human linguistic diversity, and language use and variation are still poorly understood outside of modern, industrialized states. Emerging research results have clarified the importance of studying more diverse social contexts for modeling the role of language in structuring human societies and for the study of linguistic development and change. For example, recent work on the nature of multilingualism in small-scale societies has established that its properties are much more varied than assumed by theories developed around urban settings, and work on language evolution has recognized the importance of creating models that account for the fact that, for most of human history, people lived in settings where they mostly only interacted with people that they knew. These advances in knowledge call for increased study of varied contexts of language use throughout the globe, especially given that languages are rapidly disappearing due to global-scale social forces, and for the development new research methods to help better understand the impact of this diversity on linguistic structure and development. More broadly, it is imperative to ensure that research on the world's languages is not limited to large-scale societies so that it captures the full capacities of the human mind.